Commentary on Jarl & Gerdtham (2012): Further evidence of the reversal of alcohol-related risk of oesophageal cancer after alcohol cessation


First of all, we would like to congratulate Johan & Gerdtham on their thorough meta-analysis [1], which has the potential to become a landmark study on the reversibility of the effects of alcohol.

Cancer of the oesophagus is one of the most fatal forms of cancer, with the 5-year survival rate being as low as 16.8% [2].

A number of studies, completed in various countries, have established alcohol consumption to be one of the major risk factors for oesophagus cancer [3–5], with this factor being a particularly important determinant of oesophagus cancer morbidity and mortality in countries with heavy alcohol consumption [6]. Although alcohol consumption and smoking are often correlated, an elevated risk has also been found in drinkers who do not smoke [7,8], demonstrating that the association between alcohol consumption and this cancer is independent of smoking.

Several individual epidemiological studies have examined the effect that alcohol cessation has on the risk of cancer. The studies have generally been relatively consistent in demonstrating a reduction in the risk of developing cancer of oesophagus, as well as of the oral cavity, pharynx and larynx; however, there is considerable variation in the estimates reported of both the magnitude of the risk reduction and the time lag after which the reduction occurs.

The present meta-analysis aimed to quantify the effect of alcohol cessation on the risk of developing oesophageal cancer based on nine studies (eight case–control studies and one prospective cohort study). The authors report that about 16.5 years [95% confidence interval (CI): 12.7–23.7] are required until all elevated alcohol-related risk has disappeared. The risk falls faster during the first few years of drinking cessation, and about half the elevated risk disappears after a third of the total required time.

Recently, our team conducted a pooled analysis to evaluate the temporal sequence and the strength of the association between alcohol cessation and reduction in oesophageal and head and neck cancer risks [9]. We included data from five individual studies of oesophageal cancer (all of which were also included in the analysis of Johan & Gerdtham [1]). We found that there was a 63% risk reduction for oesophageal cancer after 15 years of alcohol cessation (95% CI: 59–67%). After more than 20 years of alcohol cessation, the risks for both types of cancer were no longer statistically significantly different from their ‘never drinkers’ counterparts.

Furthermore, in our study for oesophageal cancer, we found that the risk increased significantly within the first 2 years following cessation (odds ratio: 2.50, 95% CI: 2.23–2.80). This is not surprising, as it reflects the trend that the majority of the existing individual studies observed an increase in risk following cessation due, presumably, to ‘sick quitter’ behaviour [10,11], sample characteristics (e.g. over-representation of heavy drinkers), different patterns of drinking in different countries and/or for some other reasons, as speculated in the Johan & Gerdtham paper [1]. It may be that the linear methodology they selected smoothed over these peaks, but nevertheless this is irrelevant, as there are no indications that cessation of alcohol consumption would be causally responsible for an increased risk in oesophageal cancer.

Overall, several questions, such as (i) whether a beneficial effect of drinking cessation is greater among smokers than non-smokers or (ii) what is the effect of a reduction in drinking rather than quitting, as cessation might not be an option for many drinkers, remain unanswered at this stage because of a limited number of epidemiological studies on this topic, as well as their limitations. It should also be noted that there are potentially important differences between the existing studies in terms of methodology and/or reporting; particularly, ex-drinkers were defined differently across the studies, and the age at which individuals started to drink, the duration of the habit and the reasons for stopping alcohol consumption were not reported in the majority of the studies. Furthermore, beverage- and pattern-specific risks were not controlled for, and the studies are incomparable in terms of potential confounders and the effect–measure modifiers addressed (e.g. smoking, diet, education, etc.).

Due to the above limitations of the previous studies, it may not be possible to refine the existing meta/pooled analyses [1,9] much further. More valid epidemiological studies—in different countries, with different patterns of alcohol consumption—are needed in order to advance our knowledge further on the effect of drinking cessation on the risk of oesophageal cancer.

Ultimately, accurate quantification of the beneficial effects that quitting or a reduction of drinking has on the risk of developing cancer will have important public health implications for prevention and health promotion. Fifteen to 20 years for a complete reversal of risk of oesophageal cancer to take place sounds like a long stretch; however, the effect of stopping alcohol consumption appears to emerge shortly after cessation, which is encouraging and should provide extra motivation for the promotion of this life-style change.

Declarations of interest